Is NVIDIA really unstoppable or could it be a dead man walking?
I’m posting this together with an Ed Zitron interview, these 2 combined I think gives you a very complete picture of the AI situation, and where it is probably heading.
Here the focus is on decentralized AI being more likely to be the main thing in the future for many reasons.
If you are unfamiliar with the issue, Asian Dad Energy puts up a very solid case for it IMO.Yeah, Zitron is definitely a fine pairing here.
Like I was thinking, “Yeah, no shit OpenAI uses the most expensive NIVIDIA GPUs… That’s why NVIDIA’s given OpenAI billions of dollars. OpenAI literally exists to buy expensive NVIDIA GPUs!”
Absolutely and in this case, literally actually means literally.
As with every bubble, all the people in it, think they’ve found a source of infinite growth.
This video however shows how that is very unlikely, even if use cases do explode as many speculate and is the basis for these expensive investments, they will be more likely to run on decentralized systems and not centralized as the financial bets are depending on. Nobody will control the technology, and all those data centers will at most service what is akin to an overly expensive app.I’m not particularly knowledgeable about technology or finance, but even I’ve gotten a bunch of red flags that would be easy enough to lower if they were wrong.
But one thing I’d been thinking is that AI investors are acting like they can “capture the market,” despite all the evidence pointing to that anyone with enough money can re-build everything, and that the benefits of spending more money on development are yielding smaller and smaller benefits; meaning that there is no creating a killer monopoly on AI where you can just charge whatever you want–there will always be competitors nipping at your heels, maybe with prestige models, maybe with models that aren´t as capable, but cheaper.
And the other problem is that every time you use AI, it costs as much as all the other times you’ve used AI. There’s really not an amazing benefit for scaling.
So I think what’s happening is investors are just making a lot of assumptions, thinking AI is like Facebook or Amazon or Windows–monopoly companies that can extract value from everything it touches–but it’s a lot more like Red Hat–technology that can be modestly specialized but will always be replaceable.
meaning that there is no creating a killer monopoly on AI
Absolutely, they have this idea that the market will be controlled by a few winners of a race. When the reality is that the threshold for entering will keep declining, and all the chips they’ve bought will be obsolete soon.
These Nvidia based data centers are extremely expensive to build and will probably never be profitable.thinking AI is like Facebook or Amazon or Windows–monopoly
Absolutely true, there is no critical mass for AI, if an AI works, it will work just as well with fewer users, again making it way easier to enter the market when the technology becomes cheaper. And that is true even without decentralized AI becoming dominant.
But the fact that we are able to have this conversation as non-experts (?) and nobody is really able to counter these arguments tells me that something is fundamentally wrong with the way everything is structured.
My idea is that ultimately we’re looking at the last stage of societal collapse into kleptocracy. There is no social will to stand between rich people making money and social harm, so the wealthy will cannibalize all of us until we have a revolution.
You can look at it more closely. I think it was an Adam Conover guest that quoted Hobbes on a tyrants, but included the part people leave out–that it is the tyrant who determines what is good and what is bad.
I think definitely part of of this (which you’ll hear Zitron bitch about constantly) is the total and complete failure of the business press, which is wholly captured by business. There is no feedback mechanism to stop people from making money unethically.
nobody is really able to counter these arguments
My wife made this exact observation after having seen Ed Zitron interviewed several times. No one contradicts him!
quoted Hobbes on a tyrants, but included the part people leave out–that it is the tyrant who determines what is good and what is bad.
I don’t get that part? Clearly if it’s a tyrant, the people have decided he is bad. The tyrant doesn’t get to decide his own reputation. So that makes absolutely zero sense to me. If he is good, he is a King or ruler or called something other than tyrant.
Your follow up with the press following the industry blindly, is more akin to what we call a “folke forfører” a person that somehow is seducing the people, so they lose their ability to judge.
I think Hobbes would say the only difference between a tyrant and a good king is perception, which can be gamed, and the true power of a tyrant (counter Cersei Lannister) is to define what is good and bad.
I mention it because making money and supporting people who make money seems to be “good” in this current society, and people arguing for limiting the wealth and power of a few are “bad.” That is how the anti-communist blacklist worked. You didn´t need to legislate these people out of working in hollywood, you just labelled them as bad and let society do the rest. It’s a much more efficient mechanism of control.
More trivial example: I tried shit-talking Taylor Swift at a social function and it was a huge mistake. Taylor Swift has effectively PR´d herself into being a good, relatable person, even though she is a horribly hollow shell of a human being.
(Speaking of internalizing mechanisms of control, “Burn Out Society” is pretty fantastic).
What’s the translation of folkforer literally? In english we have brown-noser, ass-kisser, boot-licker; I’ve added ball-fondler and throat-holster. In Spanish there is the chupamedias, or sock sucker, and French has le fayot, which literally translates to “bean” so I don´t know how the hell is going on there.
I don’t want to be confidently ignorant; I’m willing to hear other sides; but I’m also not willing to be buried under a bunch of information that obscures the truth. So, yeah, if your company has a $1 trillion valuation and you are actually still burning through $70 billion dollars a year and 70% of your revenue is vendors you are loaning you money… i mean, come on.
Maybe they are hoping the PC becomes so expensive that everyone is forced to run everything in the cloud. Investors might also be hoping for regulatory capture. They are saying ai is dangerou; please regulate me in to a monopoly.
I think especially that last part is 100% spot on. But I also don’t think you can stop decentralized open source AI.
Trying to help get us there faster with self-hosted pipelines: https://docs.circuitforge.tech/
This guy is AI
I’m not going to disclose how I conclude that, but there are three clues in the video that reveal it to be AI. See if you too can catch those. Don’t tell others on public forums because we all need to know how to catch AI without telling them how to correct that too.
Okay, I don’t think he is AI though. I looked at his oldest video, which admittedly is only from 7 months ago, but would have used an older and less capable slop model, and it looks very legit. Even the distortion from his high myopia lenses can be seen easily on his glasses as he moves his head left and right. The background remains entirely consistent the whole time, and even the books and pictures are consistent throughout the entire video.
It looks very real
I’m pretty sure it’s AI slop.
Can you DM me the reasons and/or timestamps :)